When I first saw Scott Sinclair’s recent ESG research indicating that 72% of large enterprises are committed to software-defined storage for their data infrastructure of the future, I thought he had to have transposed the digits: 72% is a big number; the array vendors are, in a word, entrenched, and corporate IT is already fighting multiple wars simultaneously on the security, application transition to containers, and cloud rightsizing fronts.
So with the CIO agenda for 2020 full and then some, why and when are they moving to a software-defined storage future? Scott found that 55% of enterprises have software-defined storage deployed in some capacity now, with a higher percentage indicating they plan to expand and deepen their usage of SDS in 2020. Part of this expansion hinges on adopting a multi-vendor strategy as they’ve done in all other critical parts of operations, from hardware vendors to public cloud vendors.
Our conversations with Fortune 1000 enterprises over the last two years are very consistent with Scott’s findings and reveal the main factors behind the shift to software-defined:
1. INCREASE PERFORMANCE. Software-defined systems are taking advantage of the move to 100Gb+ Ethernet and NVMe drives; throughput and latency problems are solved problems for the data center, but performance can be crippled when essential services like deduplication, compression and encryption are on. One Datera customer boasts that he cut latency by 90% over prior storage arrays by standardizing on SATA Flash and NVMe drives.
2. SIMPLIFY THROUGH AUTOMATION. Aggregating application and tenants into an enterprise cloud and increasing velocity via self-service utilization for application owners requires automation. Policy-based administration and management by application intent—where policies set thresholds like the IOPS needed for an application and the software utilizes the right resources to meet that threshold—eliminate the manual tuning required by legacy approaches, and dramatically simplify the overall environment. Automation is no longer a nice to have, but essential for operating at scale.
3. FLEXIBILITY AND CHOICE IN HARDWARE SELECTION. CIOs overwhelmingly want the ability to choose the hardware they require at the right time from a menu of choices. All too often, storage systems have limited choice to only certain types of media and to the same generation of hardware system, making system design inflexible. But SDS systems like Datera are designed to rapidly incorporate the latest new technologies and enable a broad selection of media types (NVMe, SATA Flash, HDD) as the system expands and of server hardware from a pool of vendors including DellEMC PowerEdge and HPE ProLiant spanning multiple generations, enabling choice that helps overcome the supply chain challenges the industry is currently experiencing and provide leverage to negotiate for procurement to get better pricing and delivery times.
4. SCALE FOR TODAY AND TOMORROW: Enterprises are universal in their desire to scale their systems just-in-time when they need new capacity and capabilities, and to do so granularly (node by node) from a few hundred terabytes to a hyperscale threshold of multiple petabytes. Well architectured systems not only can meet scale requirements but can utilize new capacity to expand performance, durability, and resilience as the environment grows. Datera has worked with several CIOs that embraced hyperconverged infrastructure (HCI) like VMware vSAN and Nutanix to serve a need quickly, but experienced an inability to scale beyond 10 to 20 nodes and achieve acceptable performance, which led them to Datera.
5. REDUCE OPERATING EXPENSES WITH SELF-HEALING SYSTEMS. Combining a distributed data management approach with advanced telemetry to monitor the health of each hardware node and rebalance when appropriate offers the potential to help all users with practical capacity/performance planning and best practices in real-time. One of our customers has, in his words, “eliminated hardware maintenance” by simply swapping in new hardware nodes when an individual server fails rather than take systems offline and deploy specialists to revive the faulty array. With Datera, decommissioning a node and bringing on a new node can be achieved without downtime and in a matter of minutes. And the system, rather than the administrator, rebalances the environment to re-incorporate that node and deploy the right data to it to utilize the media type.
6. INCREASE DATA AVAILABILITY. For Fortune 1000 enterprises, data availability remains paramount in system design. Properly designed software-defined systems shun typical RAID schemes and deliver availability by distributing copies of data across nodes, racks, aisles, and data centers with close attention to fault zones. This enables a higher level of availability than traditional systems. At Datera, each node gets “over-the-air” updates so there is no need to take a system offline to upgrade it, eliminating the biggest downtime culprit—planned downtime.
7. OPTIMIZE FOR CONTAINERS AND KUBERNETES. 90% of enterprises are using containers in production now, revealing cracks in the infrastructure’s ability to keep pace with the transitory nature of the applications. CIOs are actively looking for storage systems optimized for containers providers and tools like Docker, Kubernetes, and Red Hat Openshift. SDS systems can be a good choice when they can autonomous operations and optimize data placement to match the agility afforded with container deployments.
8. MINIMIZE LOCK-IN. The infrastructure industry is notorious for vendors locking customers into an inflexible architecture and an artificially limited set of hardware choices designed around their economics rather than yours. To avoid lock-in, software-defined approaches should support a wide variety of hardware profiles—different server vendors, different server models, different server generations, and a variety of different media—to keep their choice and negotiating power as high as possible. While many SDS systems are simply packaged appliances with narrow options, Datera offers a number of qualified servers from the top server manufacturers, including Dell EMC PowerEdge, HPE ProLiant DL360s and DL380s, Fujitsu Primergy MX2540s and Cisco UCS C240s.
9. REDUCE TECHNICAL DEBT. Enterprise CIOs can significantly reduce technical debt in networking and storage through SDS. Reducing technical debt is often less about total spend than it is about replacing monolithic systems with more fluid and customizable systems that change the paradigm of arrays in two ways—excessive markup and forklift upgrades. First, the raw expense of fibre channel networking and the markup vendors are exacting for all-flash and all-NVMe systems have reached a new peak, but Datera harnesses commodity servers and eliminates the three-year refresh cycle which reduces capital expense upfront and as time goes on. Datera environments are “evergreen,” meaning they are continuously refreshed without downtime.
2020 AND BEYOND
Software-defined storage for enterprise clouds crossed the chasm from early majority to late majority in 2019. As we move to 2020, the Fortune 1000 are engaging multiple software-defined vendors and moving up to SDS from HCI implementations using VMware vSAN and Nutanix that were never meant for enterprise rollouts. Scott summed it up that ESG’s “market research shows that a majority of enterprises are adopting software-defined storage technologies as part of their data strategy and Datera is emerging as one of the central players for mission-critical applications.”
Get a Consultation
Discover the advantages of enterprise software-defined storage with advanced automation: Contact us to schedule a free consultation.